Get desktop application:
View/edit binary Protocol Buffers messages
Flat run respresentation. Used for the rows of the Run Table.
Used in:
The id of the run.
The time the run was started.
The time the run ended.
The current state of the run(trial).
The tags of the associated experiment. TODO(aaron.amanuel): Create add/remove tags for runs.
The total size of checkpoints.
The count of checkpoints.
Signed searcher metrics value.
The id of external run
Trial hyperparameters.
summary metrics.
The id of the user who created the run.
Time in seconds which the run ran or has been running.
The id of the project associated with this run.
The name of the project associated with this run.
The id of the workspace associated with this run.
The name of the workspace associated with this run.
The archived status of the parent project (can be inherited from workspace).
Data related the the experiment associated with this run.
The arbitrary metadata of the run.
The archived status of this run. This is only looking at the archived status at the run level and not taking into account whether the experiment is archived or not.
Project level local id of run.
Log policy matched.
Minimal experiment object
Used in:
The id of the experiment linked to the run.
The type of searcher for the experiment.
The searcher metric name for the experiment.
Original id of a forked or continued experiment.
The id of external experiment
The resource pool the experiment was created in.
The current progress of a running experiment.
The description of the experiment.
The experiment name.
Unmanaged experiments are detached.
True if the associated experiment is a multitrial experiment
The experiment pachyderm integration config.
Flat run respresentation.
The id of the run.
The time the run was started.
The time the run ended.
The current state of the run(trial).
The tags of the associated experiment.
The total size of checkpoints.
The count of checkpoints.
Signed searcher metrics value.
The id of external run
Trial hyperparameters.
summary metrics.
The id of the user who created the run.
Time in seconds which the run ran or has been running.
The id of the project associated with this run.
The name of the project associated with this run.
The id of the workspace associated with this run.
The name of the workspace associated with this run.
The archived status of the parent project (can be inherited from workspace).
The arbitrary metadata of the run.
The archived status of this run. This is only looking at the archived status at the run level and not taking into account whether the search is archived or not.
The id of the search linked to the run.
The type of searcher for the search.
The searcher metric name for the search.
Original id of a forked or continued search.
The id of external search
The resource pool the search was created in.
The current progress of a running search.
The description of the search.
The search name.
Unmanaged searchs are detached.
True if the associated search is a multitrial search
The search pachyderm integration config.
Number times the trial restarted.
The current processed batches.
Best validation.
Latest validation.
Best checkpoint.
The last reported state of the trial runner (harness code).
The wall clock time is all active time of the cluster for the trial, inclusive of everything (restarts, initiailization, etc), in seconds.
UUID of checkpoint that this trial started from.
Task IDs of tasks associated with this trial. Length of task_ids will always be greater or equal to one when TaskID is sent. For example CompareTrial we will send a reduced Trial object, without TaskID or TaskIDs fileld in. The first element of task_ids will be the same as task_id. task_ids is sorted ascending by task_run_id.
Number of days to retain logs for.